Edge side terminal security threat analysis and evaluation method
By extracting and encoding the operating status, network traffic and user behavior data of edge-side terminals, and combining with Bayesian probability networks for correlation analysis, security threat posterior representation vector is generated, solving the accuracy of security threat detection on edge-side terminals The problem of insufficient improvements in dynamic adaptability and security guarantee capabilities.
Patent Information
- Application Number
- CN202411145325.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-20
- Publication Date
- 2024-12-06
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The prior art is difficult to effectively deal with the dynamically changing attack patterns and hidden malicious behaviors faced by edge-side terminals, resulting in insufficient accuracy and timeliness of security threat detection.
By obtaining the operating status data, network traffic values and user behavior data of edge-side devices, timing feature extraction and encoding are performed, and feature interaction correlation analysis is performed in combination with Bayesian probability network to generate security threat posterior timing representation vectors, which are used to determine edges Is there any security threat to the side terminal?
It improves the dynamic adaptability and accuracy of security threat detection, effectively responds to complex environments and unknown attacks, and enhances the security guarantee capabilities of edge computing environments.
Smart Images

Figure CN119089449A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of intelligent analysis and evaluation, and specifically to a method for analyzing and evaluating security threats on edge terminals. Background Art
[0002] With the rapid development of technologies such as the Internet of Things (IoT), big data, and cloud computing, edge computing, as an emerging computing model, is gradually becoming a key technology to support the Internet of Everything, real-time data processing, and efficient service response. Edge computing effectively reduces network bandwidth requirements, improves response speed, and reduces latency by moving computing and data storage capabilities to the edge of the network, that is, processing at the user or device end. However, the widespread application of this computing model also brings many security threats and challenges.
[0003] At present, the security assessment of edge terminals mostly relies on traditional static security policies and signature matching methods, which are often unable to effectively deal with the dynamically changing attack modes and hidden malicious behaviors faced by edge terminals, resulting in insufficient accuracy and timeliness of security threat detection. Therefore, an optimized edge terminal security threat analysis and assessment method is expected. Summary of the invention
[0004] This application is made in consideration of the above problems. One purpose of this application is to provide a method for analyzing and evaluating security threats of edge terminals.
[0005] The embodiment of the present application provides a method for analyzing and assessing security threats of edge terminals, which includes:
[0006] Obtaining the time series of the running status data of the edge side device, the time series of the network traffic value, and the time series of the user behavior data, wherein the running status data includes the device temperature, device voltage, fan speed, memory percentage, and CPU usage;
[0007] Performing running state time series feature extraction and feature enhancement processing on the time series of the running state data to obtain an enhanced running state time series associated feature vector;
[0008] Performing time series association coding on the time series of the network traffic value and the time series of the user behavior data respectively to obtain a network traffic local time series association implicit feature vector and a user behavior time series semantic coding feature vector;
[0009] Performing feature interaction correlation analysis on the enhanced operation state time series correlation feature vector, the network traffic local time series correlation implicit feature vector, and the user behavior time series semantic coding feature vector to obtain a security threat posterior time series representation vector;
[0010] Based on the security threat a posteriori time series representation vector, it is determined whether there is a security threat to the edge side terminal.
[0011] For example, according to the edge terminal security threat analysis and assessment method of an embodiment of the present application, the time series of the running state data is subjected to running state time series feature extraction and feature enhancement processing to obtain an enhanced running state time series associated feature vector, including:
[0012] Performing full-time domain collaborative correlation analysis on the time series of the operating status data to obtain an operating status time series correlation feature graph;
[0013] Inputting the running state temporal correlation feature map into a feature attention selection and enhancement module based on a compression-inhibition structure to obtain an enhanced running state temporal correlation feature map;
[0014] The global mean vector of the enhanced operating state time series association feature graph is calculated to obtain the enhanced operating state time series association feature vector.
[0015] For example, according to the edge terminal security threat analysis and assessment method of an embodiment of the present application, the time series of the operating status data is subjected to full-time domain collaborative correlation analysis to obtain an operating status time series correlation feature graph, including:
[0016] The time series of the operating status data is arranged into an operating status full-time domain collaborative correlation matrix according to the parameter sample dimension and the time dimension, and then the operating status time series pattern feature extractor based on the void convolutional neural network model is used to obtain the operating status time series correlation feature graph.
[0017] For example, according to the edge terminal security threat analysis and assessment method of an embodiment of the present application, the operating state time series correlation feature map is input into a feature attention selection enhancement module based on a compression-suppression structure to obtain an enhanced operating state time series correlation feature map, including:
[0018] Calculating the global mean of each feature matrix along the channel dimension of the running state time series associated feature graph to obtain a running state time series feature compression information representation vector;
[0019] Performing multi-scale feature extraction on the running state time series feature compression information representation vector to obtain a multi-scale associated feature vector of the running state time series feature compression information;
[0020] Using a Sigmoid function, a normalization operation is performed on the multi-scale associated feature vector of the running state time series feature compression information to obtain a feature weight vector representing the running state time series feature;
[0021] Based on the running state time series feature representation feature weight vector, feature amplification and suppression operations are performed on the running state time series association feature graph to obtain the enhanced running state time series association feature graph.
[0022] For example, according to the edge terminal security threat analysis and assessment method of an embodiment of the present application, multi-scale feature extraction is performed on the running state time series feature compression information representation vector to obtain a multi-scale associated feature vector of the running state time series feature compression information, including:
[0023] Performing one-dimensional convolution coding on the running state time series feature compression information representation vector to obtain a correlation representation feature vector between the running state time series feature compression information;
[0024] Cascading the running state time series feature compression information representation vector and the running state time series feature compression information association representation feature vector to obtain the running state time series feature compression information multi-scale representation vector;
[0025] The multi-scale representation vector of the compressed information of the operating status time series feature is input into a compressed information feature extraction module based on a multi-layer perceptron model to obtain the multi-scale associated feature vector of the compressed information of the operating status time series feature, wherein the multi-layer perceptron model includes two fully connected layers and a SiLU activation function.
[0026] For example, according to the edge terminal security threat analysis and assessment method of an embodiment of the present application, based on the running state time series feature representation feature weight vector, the running state time series association feature graph is subjected to feature amplification and suppression operations to obtain the enhanced running state time series association feature graph, including:
[0027] The enhanced running state time series association feature graph is obtained by calculating and multiplying the corresponding feature values of the feature weight vector representing the running state time series feature with the feature matrices along the channel dimension of the running state time series association feature graph.
[0028] For example, according to the edge terminal security threat analysis and assessment method of an embodiment of the present application, the time series of the network traffic value and the time series of the user behavior data are respectively subjected to time series association coding to obtain a network traffic local time series association implicit feature vector and a user behavior time series semantic coding feature vector, including:
[0029] Inputting the time series of the network traffic value into a sequence encoder based on an RNN model to obtain a local temporal correlation implicit feature vector of the network traffic;
[0030] The time series of the user behavior data is input into a temporal context encoder including a user behavior embedding encoder to obtain a temporal semantic encoding feature vector of the user behavior.
[0031] For example, according to the edge terminal security threat analysis and assessment method of an embodiment of the present application, the enhanced operating state time series correlation feature vector, the network traffic local time series correlation implicit feature vector and the user behavior time series semantic coding feature vector are subjected to feature interaction correlation analysis to obtain a security threat posterior time series representation vector, including:
[0032] The enhanced operating state time series association feature vector, the network traffic local time series association implicit feature vector and the user behavior time series semantic coding feature vector are input into a security threat posterior analyzer based on a Bayesian probability network to obtain the security threat posterior time series representation vector.
[0033] For example, according to the edge terminal security threat analysis and assessment method of an embodiment of the present application, determining whether there is a security threat to the edge terminal based on the security threat posterior time series representation vector includes:
[0034] The security threat posterior time series representation vector is input into a classifier-based threat assessment result generator to obtain an assessment result, and the assessment result is used to indicate whether there is a security threat.
[0035] According to the edge terminal security threat analysis and assessment method of the embodiment of the present application, it can improve the dynamic adaptability and accuracy of security threat detection, effectively respond to the challenges of various complex environments and unknown attacks, thereby enhancing the security protection capabilities of the edge computing environment. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings of the embodiments of the present application are briefly introduced below. Obviously, the drawings described below only relate to some embodiments of the present application, and are not intended to limit the present application.
[0037] Figure 1 A schematic diagram of the application architecture of the edge terminal security threat analysis and assessment method in an embodiment of the present application is shown;
[0038] Figure 2 A flowchart of a method for analyzing and assessing security threats of edge terminals in an embodiment of the present application is shown;
[0039] Figure 3 A flowchart of sub-step S520 of the edge terminal security threat analysis and assessment method in an embodiment of the present application is shown;
[0040] Figure 4A schematic diagram of the structure of the edge terminal security threat analysis and assessment system in an embodiment of the present application is shown;
[0041] Figure 5 An application scenario diagram of the edge terminal security threat analysis and assessment method in an embodiment of the present application is shown; and
[0042] Figure 6 The security model structure of the power Internet of Things edge-side threat capture and adaptive response linkage in an embodiment of the present application is shown. DETAILED DESCRIPTION
[0043] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work also fall within the scope of protection of the present application.
[0044] The terms used in this specification are those common terms currently widely used in the art in consideration of the functions of the present application, but these terms may vary according to the intentions of those skilled in the art, precedents, or new technologies in the art. In addition, specific terms may be selected, and in this case, their detailed meanings will be described in the detailed description of the present application. Therefore, the terms used in the specification should not be understood as simple names, but rather as a general description based on the meaning of the terms and the present application.
[0045] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are only illustrative, and different aspects of the system and method can use different modules.
[0046] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps may be processed in reverse order or simultaneously as required. At the same time, other operations may also be added to these processes, or a certain step or several steps of operations may be removed from these processes.
[0047] Figure 1 A schematic diagram of the application architecture of the edge terminal security threat analysis and assessment method in an embodiment of the present application is shown, including a server 100 and a terminal device 200.
[0048] The terminal device 200 and the server 100 can be connected via the Internet to achieve mutual communication. Optionally, the above-mentioned Internet uses standard communication technology and / or protocol. The Internet is usually the Internet, but it can also be any network, including but not limited to any combination of a local area network (LAN), a metropolitan area network (MAN), a wide area network (WAN), a mobile, wired or wireless network, a dedicated network or a virtual private network. In some embodiments, the data exchanged through the network is represented by technologies and / or formats including Hyper Text Markup Language (HTML), Extensible Markup Language (XML), etc. In addition, conventional encryption technologies such as Secure Socket Layer (SSL), Transport Layer Security (TLS), Virtual Private Network (VPN), Internet Protocol Security (IPsec) can also be used to encrypt all or some links. In other embodiments, customized and / or dedicated data communication technologies can also be used to replace or supplement the above-mentioned data communication technologies.
[0049] The server 100 can provide various network services for the terminal device 200, wherein the server 100 can be a single server, a server cluster consisting of several servers, or a cloud computing center. Specifically, the server 100 may include a processor 110 (Center Processing Unit, CPU), a memory 120, an input device 130, and an output device 140, etc. The input device 130 may include a keyboard, a mouse, a touch screen, etc., and the output device 140 may include a display device, such as a liquid crystal display (Liquid Crystal Display, LCD), a cathode ray tube (Cathode Ray Tube, CRT), etc.
[0050] The memory 120 may include a read-only memory (ROM) and a random access memory (RAM), and provides the processor 110 with program instructions and data stored in the memory 120. In an embodiment of the present application, the memory 120 may be used to store the program of the edge terminal security threat analysis and assessment method in an embodiment of the present application.
[0051] The processor 110 calls the program instructions stored in the memory 120, and the processor 110 is used to execute the steps of any one of the edge terminal security threat analysis and assessment methods in the embodiments of the present application according to the obtained program instructions.
[0052] In addition, the application architecture diagram in the embodiment of the present application is intended to more clearly illustrate the technical solution in the embodiment of the present application, and does not constitute a limitation on the technical solution provided in the embodiment of the present application. Of course, for other application architectures and business applications, the technical solution provided in the embodiment of the present application is also applicable to similar problems.
[0053] The following is a non-restrictive description of the edge terminal security threat analysis and assessment method provided according to at least one embodiment of the present application through several examples or embodiments. As described below, different features in these specific examples or embodiments can be combined with each other without conflicting with each other to obtain new examples or embodiments, and these new examples or embodiments also fall within the scope of protection of this application.
[0054] In response to the above technical problems, the technical concept of this application is to collect and analyze real-time data of edge devices, network traffic and user behavior, respectively dig out the time-series change patterns of the operating status of edge devices, network traffic and user behavior, and then construct the security threat feature representation of edge terminals based on the feature association between the three, so as to realize intelligent judgment of security threats to edge terminals. In this way, the dynamic adaptability and accuracy of security threat detection can be improved, and the challenges of various complex environments and unknown attacks can be effectively responded to, thereby enhancing the security protection capabilities of the edge computing environment.
[0055] Based on this, Figure 2 The flowchart of the edge terminal security threat analysis and assessment method in the embodiment of the present application is shown. For example, the edge terminal security threat analysis and assessment method can be executed by a server, which can be Figure 1 The server 100 shown in FIG. Figure 2As shown, according to the edge terminal security threat analysis and assessment method of the embodiment of the present application, the steps include: S510, obtaining the time series of the running status data of the edge device, the time series of the network traffic value and the time series of the user behavior data, wherein the running status data includes the device temperature, the device voltage, the fan speed, the memory share and the CPU usage; S520, performing running status time series feature extraction and feature enhancement processing on the time series of the running status data to obtain an enhanced running status time series association feature vector; S530, performing time series association coding on the time series of the network traffic value and the time series of the user behavior data respectively to obtain a network traffic local time series association implicit feature vector and a user behavior time series semantic coding feature vector; S540, performing feature interaction correlation analysis on the enhanced running status time series association feature vector, the network traffic local time series association implicit feature vector and the user behavior time series semantic coding feature vector to obtain a security threat posterior time series representation vector; S550, based on the security threat posterior time series representation vector, determining whether there is a security threat on the edge terminal.
[0056] Specifically, in the technical solution of the present application, first, the time series of the running status data of the edge side device, the time series of the network traffic value and the time series of the user behavior data are obtained, wherein the running status data includes device temperature, device voltage, fan speed, memory share and CPU usage. It should be understood that the device temperature, device voltage and fan speed of the edge side device directly reflect the physical health status and energy supply of the device, while the memory share and CPU usage reflect the load and computing activity of the device. Excessive resource occupancy may be a sign of malware activity, denial of service attack or resource exhaustion attack. At the same time, network traffic is an important indicator of the interaction between the edge side device and the outside world, and abnormal traffic patterns may indicate potential intrusion or malicious activity. The user behavior data contains information such as the user's login time, operation sequence, access resources, etc. Abnormal user behavior may be a manifestation of account theft, internal threat or social engineering attack. Through in-depth analysis of the above multi-dimensional data, more abundant security threat clues can be captured and the security status of the edge side terminal can be more comprehensively understood.
[0057] Next, considering the multi-dimensional characteristics of the edge device operation status data, as well as the possible correlation effects and interactions between different operation status parameters, in order to deeply understand the synergistic effects of various state parameters in the operation status data in the time dimension, the time series of the operation status data is further arranged according to the parameter sample dimension and the time dimension into an operation status full-time domain synergistic correlation matrix to establish potential correlations between parameters. Then, the operation status time series pattern feature extractor based on the dilated convolutional neural network model is used to extract features from the operation status full-time domain synergistic correlation matrix, and the synergistic correlation change pattern of each state parameter over time is captured through the convolution operation. The receptive field is expanded through the dilated structure, and the long-distance time series dependency of the operation status parameters is effectively captured. The dynamic evolution process of the equipment operation status is deeply understood, thereby generating an operation status time series correlation feature map.
[0058] Then, considering that in the operating state time series correlation feature map, there may be mutations or abnormal patterns of certain key state parameters that indicate security threats, while other parameters may be relatively flat, with weak indication of security threats, or have low correlation with the current security state. Therefore, in the technical solution of the present application, a feature attention selection and enhancement module based on a compression-suppression structure is further introduced to selectively focus on the feature information in the operating state time series correlation feature map. Specifically, the feature attention selection and enhancement module is based on an adaptive compression-suppression mechanism. First, it performs feature compression on the operating state time series correlation feature map to reduce redundant information and obtain global channel information. Then, it performs multi-scale feature extraction on its global channel information to capture the information association between the feature channels of the operating state time series correlation feature map, and generates corresponding attention weights accordingly. Each feature channel of the operating state time series correlation feature map is selectively suppressed and amplified, thereby obtaining an enhanced operating state time series correlation feature map.
[0059] Then, the global mean vector of the enhanced operating status timing association feature graph is further calculated to retain the key timing patterns of the enhanced operating status timing association feature graph and filter out redundant information such as background noise to obtain the global timing representation of the operating status of the edge-side device, thereby obtaining the enhanced operating status timing association feature vector and further improving the pertinence and effectiveness of the feature representation.
[0060] Accordingly, in step S520, Figure 3As shown, the time series of the running status data is subjected to running status temporal feature extraction and feature enhancement processing to obtain an enhanced running status temporal association feature vector, including: S521, performing full-time domain collaborative association analysis on the time series of the running status data to obtain an running status temporal association feature map; S522, inputting the running status temporal association feature map into a feature attention selection enhancement module based on a compression-suppression structure to obtain an enhanced running status temporal association feature map; S523, calculating the global mean vector of the enhanced running status temporal association feature map to obtain the enhanced running status temporal association feature vector.
[0061] Among them, in step S521, a full-time domain collaborative correlation analysis is performed on the time series of the operating status data to obtain an operating status time series correlation feature diagram, including: arranging the time series of the operating status data into an operating status full-time domain collaborative correlation matrix according to the parameter sample dimension and the time dimension, and then passing the operating status time series pattern feature extractor based on the void convolutional neural network model to obtain the operating status time series correlation feature diagram.
[0062] Among them, in step S522, the running state time series association feature map is input into the feature attention selection and enhancement module based on the compression-suppression structure to obtain the enhanced running state time series association feature map, including: calculating the global mean of each feature matrix of the running state time series association feature map along the channel dimension to obtain the running state time series feature compression information representation vector; performing multi-scale feature extraction on the running state time series feature compression information representation vector to obtain the running state time series feature compression information multi-scale association feature vector; using the Sigmoid function to perform a normalization operation on the running state time series feature compression information multi-scale association feature vector to obtain the running state time series feature representation feature weight vector; based on the running state time series feature representation feature weight vector, performing feature amplification and suppression operations on the running state time series association feature map to obtain the enhanced running state time series association feature map.
[0063] Specifically, multi-scale feature extraction is performed on the running state time series feature compression information representation vector to obtain a multi-scale associated feature vector of the running state time series feature compression information, including: performing one-dimensional convolution encoding on the running state time series feature compression information representation vector to obtain a running state time series feature compression information correlation representation feature vector; cascading the running state time series feature compression information representation vector and the running state time series feature compression information correlation representation feature vector to obtain a running state time series feature compression information multi-scale representation vector; inputting the running state time series feature compression information multi-scale representation vector into a compression information feature extraction module based on a multi-layer perceptron model to obtain the running state time series feature compression information multi-scale associated feature vector, wherein the multi-layer perceptron model comprises two fully connected layers and a SiLU activation function.
[0064] Specifically, based on the feature weight vector representing the operating state timing characteristics, feature amplification and suppression operations are performed on the operating state timing association feature graph to obtain the enhanced operating state timing association feature graph, including: calculating the corresponding multiplication of each eigenvalue of the feature weight vector representing the operating state timing characteristics and each feature matrix along the channel dimension of the operating state timing association feature graph to obtain the enhanced operating state timing association feature graph.
[0065] In a specific example, the operation state time series association feature map is input into a feature attention selection and enhancement module based on a compression-suppression structure to obtain an enhanced operation state time series association feature map, including: performing attention selection on the operation state time series association feature map using the following feature attention selection and enhancement formula to obtain the enhanced operation state time series association feature map, wherein the feature attention selection and enhancement formula is:
[0066]
[0067] V s =concat(V s1 ,Conv 1d (V s1 ))
[0068]
[0069] Wherein, F represents the operating state time series correlation characteristic diagram, The xth time series correlation characteristic diagram of the operating state is represented o The coordinates in the channel are (i o , j o ), H represents the height of the operating state time series correlation feature graph, W represents the width of the operating state time series correlation feature graph, is the xth vector representing the compressed information of the operating status time series feature o The characteristic value, V s1 Represents the compressed information representation vector of the operating state time series feature, Conv 1d (·) represents a one-dimensional convolution operation, concat(·,·) represents a cascade operation, V s represents the multi-scale representation vector of the compressed information of the time series feature of the running status, MLP represents the multi-layer perceptron, sigmoid represents the sigmoid function, represents the channel-by-channel dot product, and F′ represents the enhanced operating state time series correlation feature map.
[0070] Secondly, considering that information such as periodic changes, burst traffic, and abnormal patterns in network traffic data plays an important indicative role in security threat identification, in the technical solution of the present application, a sequence encoder based on an RNN model is further used to perform time series encoding on the time series of the network traffic value, so as to utilize the cyclic recursive structure of the RNN model to capture the temporal dependency of the traffic data, and effectively extract the implicit temporal feature patterns in the network traffic data, thereby obtaining the implicit feature vector of the local temporal association of the network traffic, so as to reveal possible network attacks or abnormal communication behaviors.
[0071] At the same time, in order to fully understand the temporal dynamic evolution process of user behavior, a temporal context encoder including a user behavior embedding encoder is further used to perform temporal analysis on the time series of the user behavior data. That is, first, each user behavior data in the time series of the user behavior data is converted into an embedded vector form through text embedding coding technology to capture the intrinsic semantic characteristics of user behavior. Furthermore, a temporal context encoder is used to capture the dependency of user behavior in the time dimension, that is, the before-and-after correlation and trend changes of user behavior, to understand the dynamic context information of user behavior, and thus generate a temporal semantic encoding feature vector of user behavior. In an embodiment of the present application, the temporal context encoder can be based on structures such as a long short-term memory network (LSTM), a gated recurrent unit (GRU) or a Transformer.
[0072] Correspondingly, in step S530, the time series of the network traffic value and the time series of the user behavior data are respectively subjected to time series association encoding to obtain a local time series association implicit feature vector of the network traffic and a user behavior time series semantic encoding feature vector, including: inputting the time series of the network traffic value into a sequence encoder based on an RNN model to obtain the local time series association implicit feature vector of the network traffic; inputting the time series of the user behavior data into a time series context encoder including a user behavior embedding encoder to obtain the user behavior time series semantic encoding feature vector.
[0073] Furthermore, in order to comprehensively consider the information correlation between the operating status of the edge side device, network traffic and user behavior, the enhanced operating status time series correlation feature vector, the network traffic local time series correlation implicit feature vector and the user behavior time series semantic coding feature vector are further input into the security threat posterior analyzer based on the Bayesian probability network for correlation analysis. According to the causal logical relationship between the operating status of the edge side device, network traffic and user behavior, the user behavior time series semantic coding feature vector is used as the prior probability, the enhanced operating status time series correlation feature vector is used as the conditional probability, and the network traffic local time series correlation implicit feature vector is used as evidence. The conditional probability modeling capability of the Bayesian network is used to deeply explore the intrinsic connection and mutual influence between the three, and perform posterior probability reasoning of the edge side device security threat to obtain the security threat posterior time series representation vector.
[0074] Correspondingly, in step S540, feature interaction correlation analysis is performed on the enhanced operating state timing association feature vector, the network traffic local timing association implicit feature vector and the user behavior timing semantic coding feature vector to obtain a posterior security threat timing representation vector, including: inputting the enhanced operating state timing association feature vector, the network traffic local timing association implicit feature vector and the user behavior timing semantic coding feature vector into a security threat posterior analyzer based on a Bayesian probability network to obtain the security threat posterior timing representation vector.
[0075] Next, the security threat posterior time series representation vector is input into a classifier-based threat assessment result generator to obtain an assessment result, which is used to indicate whether there is a security threat. That is, a classification algorithm is used to perform feature analysis and classification discrimination on the security threat posterior time series representation vector, and the security threat posterior time series representation vector is mapped to a preset security threat assessment category space to generate a corresponding assessment result to indicate whether there is a security threat to the edge terminal. In this way, real-time, dynamic and comprehensive assessment of edge terminal security threats can be achieved, the accuracy and timeliness of security threat detection can be improved, and various security challenges in the edge computing environment can be effectively addressed.
[0076] Accordingly, in step S550, based on the security threat posterior time series representation vector, it is determined whether there is a security threat in the edge side terminal, including: inputting the security threat posterior time series representation vector into a classifier-based threat assessment result generator to obtain an assessment result, and the assessment result is used to indicate whether there is a security threat.
[0077] It should be understood that the role of the classifier is to use the given categories and known training data to learn classification rules and classifiers, and then classify (or predict) unknown data. Logistic regression and SVM are often used to solve binary classification problems. For multi-class classification problems, logistic regression or SVM can also be used, but multiple binary classifications are required to form a multi-classification. However, this is prone to errors and is not efficient. Commonly used multi-classification methods include the Softmax classification function.
[0078] Preferably, the security threat posterior temporal representation vector is input into a threat assessment result generator based on a classifier to obtain an assessment result, including: calculating the first norm and second norm of the mean vector of the enhanced operating state temporal association feature vector, the network traffic local temporal association implicit feature vector and the user behavior temporal semantic coding feature vector; calculating the weighted sum of the inverse of the square root of the second norm and the first norm, and performing point multiplication with the point addition sum vector of the enhanced operating state temporal association feature vector, the network traffic local temporal association implicit feature vector and the user behavior temporal semantic coding feature vector to obtain a first security threat posterior temporal correction subvector; and converting the enhanced operating state temporal association feature vector, The dot product vector of the local temporal association implicit feature vector of the network traffic and the user behavior temporal semantic coding feature vector is multiplied by the square root of the length of the security threat posterior temporal representation vector to obtain a second security threat posterior temporal correction subvector; the weighted sum of the first security threat posterior temporal correction subvector and the second security threat posterior temporal correction subvector is calculated to obtain a security threat posterior temporal correction vector; the security threat posterior temporal correction vector is dot-multiplied with the security threat posterior temporal representation vector to obtain a corrected security threat posterior temporal representation vector; and the corrected security threat posterior temporal representation vector is input into a classifier-based threat assessment result generator to obtain an assessment result.
[0079] The calculation process of the security threat posterior time series correction vector is expressed as follows:
[0080]
[0081] Among them, V i (i=1-3) are respectively the enhanced running state temporal correlation feature vector, the network traffic local temporal correlation implicit feature vector and the user behavior temporal semantic coding feature vector, V μis the mean vector of the enhanced operating state temporal association feature vector, the network traffic local temporal association implicit feature vector, and the user behavior temporal semantic encoding feature vector, ||·||1 represents the first norm of the feature vector, ||·||2 represents the second norm of the feature vector, L is the length of the security threat posterior temporal representation vector, and α and β are weighted sums of weights as hyperparameters, represents point-wise addition, ⊙ represents point-wise multiplication, and V′ represents the security threat posterior temporal correction vector.
[0082] Here, considering that the enhanced operating status time series association feature vector, the network traffic local time series association implicit feature vector and the user behavior time series semantic encoding feature vector respectively represent the time series association characteristics of the operating status data, network traffic value and user behavior data of the edge side device, when performing posterior inference of sequence features based on the Bayesian probability network, it is expected to further compensate for the extreme category deviation caused by the imbalance of correspondence ratio under the fine-grained distribution of the time series, thereby improving the accuracy of the classification results.
[0083] Therefore, based on the constrained representation of the structured foreground and background distinction of the supermanifold by the mean vector norm between the enhanced operating status timing association feature vector, the local timing association implicit feature vector of the network traffic and the user behavior timing semantic coding feature vector, the feature-level key correspondence between the feature vectors of the enhanced operating status timing association feature vector, the local timing association implicit feature vector of the network traffic and the user behavior timing semantic coding feature vector is modeled, and the global correlation relationship between the corresponding features is adjusted, so as to make positive fine-grained corresponding suggestions by controlling the unbalanced proportion between the corresponding feature values of the enhanced operating status timing association feature vector, the local timing association implicit feature vector of the network traffic and the user behavior timing semantic coding feature vector, so as to avoid extreme imbalance of categories between the feature vectors by focusing on the focus.
[0084] In this way, by correcting the security threat posterior time series representation vector through the security threat posterior time series correction vector, the accuracy of the assessment result obtained by inputting the security threat posterior time series representation vector into the threat assessment result generator based on the classifier can be improved.
[0085] Based on the above embodiments, see Figure 4, which is a structural diagram of an edge terminal security threat analysis and assessment system 800 in an embodiment of the present application. The edge terminal security threat analysis and assessment system 800 includes: a data acquisition module 810, which is used to obtain the time series of running status data of the edge device, the time series of network traffic values, and the time series of user behavior data, wherein the running status data includes device temperature, device voltage, fan speed, memory share, and CPU usage; an extraction and enhancement module 820, which is used to extract running status time series features and enhance features of the time series of the running status data to obtain an enhanced running status time series correlation feature vector; a time series association coding module 830, which is used to perform time series association coding on the time series of the network traffic value and the time series of the user behavior data respectively to obtain a network traffic local time series correlation implicit feature vector and a user behavior time series semantic coding feature vector; an interactive correlation analysis module 840, which is used to perform feature interactive correlation analysis on the enhanced running status time series correlation feature vector, the network traffic local time series correlation implicit feature vector, and the user behavior time series semantic coding feature vector to obtain a security threat posterior time series representation vector; a security analysis module 850, which is used to determine whether there is a security threat to the edge terminal based on the security threat posterior time series representation vector.
[0086] Here, those skilled in the art can understand that the specific functions and operations of each module in the edge terminal security threat analysis and assessment system 800 have been described in detail above. Figure 2 to Figure 3 The edge-side terminal security threat analysis and assessment method has been introduced in detail, and therefore, its repeated description will be omitted.
[0087] Figure 5 FIG. 1 is an application scenario diagram of the edge terminal security threat analysis and assessment method according to an embodiment of the present application. Figure 5 As shown, in this application scenario, first, the time series of the running status data of the edge side device is obtained (for example, Figure 5 D1 as shown in the figure), time series of network traffic values (e.g., Figure 5 D2 as shown in the figure) and time series of user behavior data (e.g., Figure 5 D3 shown in the figure), then, the time series of the running status data, the time series of the network traffic value and the time series of the user behavior data are input to a server (for example, Figure 5 In S) shown in , the server is able to use the edge-side terminal security threat analysis and assessment algorithm to process the time series of the operating status data, the time series of the network traffic value, and the time series of the user behavior data to obtain an assessment result indicating whether there is a security threat.
[0088] Furthermore, the present application also provides a security model for studying the linkage between threat capture and adaptive response at the edge of the power Internet of Things. The security attack and defense of the power Internet of Things is changing rapidly. In order to solve the problems of high false alarm rate of threat alarms at the edge, weak response and disposal capabilities, poor linkage effect, and low utilization rate of captured threat data, a security model for the linkage between threat capture and adaptive response at the edge of the power Internet of Things is constructed to achieve comprehensive integration of security data at the edge of the power Internet of Things, in-depth threat detection, rapid disposal of security incidents, and global linkage of security defense. The model mainly includes multi-dimensional data collection, edge threat capture, edge threat detection, adaptive response linkage and other main links. In the multi-dimensional data collection stage, real-time security data collection of edge-side terminal devices, hardware, software, and user behaviors is performed to extract asset information, log information, threat intelligence and other data, providing a basis for the implementation of threat detection and threat capture in the next stage. In the threat detection stage, security data, network traffic data and threat intelligence are combined to intelligently analyze external attack behaviors, internal abnormal behaviors, and shortcomings in the configuration of own security protection, so as to achieve "precise perception" of edge-side threats. In the edge-side threat capture stage, camouflaged deception baits are implanted in new edge-side businesses to trap threats, analyze threats, and mine attack indicators, attack intentions, attack tools and attack behaviors to guide security policy deployment and security management and control implementation. In the adaptive response linkage stage, according to the results of edge-side threat detection and capture, security defense strategies are intelligently and adaptively optimized, and existing security mechanisms are linked to achieve rapid containment of edge-side threats.
[0089] Further, Figure 6 The security model structure of threat capture and adaptive response linkage at the edge of the power Internet of Things is illustrated.
[0090] Based on the above embodiments, another exemplary embodiment of an electronic device is also provided in the embodiments of the present application. In some possible implementations, the electronic device in the embodiments of the present application may include a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor can implement the steps of the edge terminal security threat analysis and assessment method in the above embodiments when executing the program.
[0091] For example, in the case of electronic equipment Figure 1 Taking the server 100 in the example as an example, the processor in the electronic device is the processor 110 in the server 100, and the memory in the electronic device is the memory 120 in the server 100.
[0092] The embodiment of the present application also provides a computer-readable storage medium, on which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, the edge terminal security threat analysis and assessment method according to the embodiment of the present application described with reference to the above figures can be executed. The computer-readable storage medium includes, but is not limited to, for example, volatile memory and / or non-volatile memory. The volatile memory may, for example, include a random access memory (RAM) and / or a cache memory (cache), etc. The non-volatile memory may, for example, include a read-only memory (ROM), a hard disk, a flash memory, etc.
[0093] The embodiment of the present application also provides a computer program product or a computer program, which includes computer executable instructions, and the computer executable instructions are stored in a computer-readable storage medium. The processor of the computer device reads the computer executable instructions from the computer-readable storage medium, and the processor executes the computer executable instructions, so that the computer device performs the edge terminal security threat analysis and assessment method according to the embodiment of the present application.
[0094] Those skilled in the art will appreciate that the contents disclosed in this application may be subject to various variations and improvements. For example, the various devices or components described above may be implemented by hardware, or by software, firmware, or a combination of some or all of the three.
[0095] In addition, although the present application makes various references to certain units in the system according to embodiments of the present application, any number of different units can be used and run on the client and / or server. The units are only illustrative, and different aspects of the system and method can use different units.
[0096] Those skilled in the art will appreciate that all or part of the steps in the above method can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, such as a read-only memory, a disk or an optical disk. Optionally, all or part of the steps in the above embodiment can also be implemented using one or more integrated circuits. Accordingly, each module / unit in the above embodiment can be implemented in the form of hardware or in the form of software function modules. The present application is not limited to any particular form of combination of hardware and software.
[0097] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by those of ordinary skill in the art to which this application belongs. It should also be understood that terms such as those defined in common dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an idealized or extremely formal sense, unless explicitly defined as such herein.
[0098] The above is an explanation of the present application and should not be considered as limiting thereof. Although several exemplary embodiments of the present application have been described, those skilled in the art will readily appreciate that many modifications may be made to the exemplary embodiments without departing from the novel teachings and advantages of the present application.
Claims
1. A method for analyzing and assessing security threats of edge terminals, characterized in that: include: Obtaining the time series of the running status data of the edge side device, the time series of the network traffic value, and the time series of the user behavior data, wherein the running status data includes the device temperature, device voltage, fan speed, memory percentage, and CPU usage; Performing running state time series feature extraction and feature enhancement processing on the time series of the running state data to obtain an enhanced running state time series associated feature vector; Performing time series association coding on the time series of the network traffic value and the time series of the user behavior data respectively to obtain a network traffic local time series association implicit feature vector and a user behavior time series semantic coding feature vector; Performing feature interaction correlation analysis on the enhanced operation state time series correlation feature vector, the network traffic local time series correlation implicit feature vector, and the user behavior time series semantic coding feature vector to obtain a security threat posterior time series representation vector; Based on the security threat a posteriori time series representation vector, it is determined whether there is a security threat to the edge side terminal.
2. The edge terminal security threat analysis and assessment method according to claim 1 is characterized in that: Extracting and enhancing the time series of the running state data to obtain an enhanced running state time series associated feature vector includes: Performing full-time domain collaborative correlation analysis on the time series of the operating status data to obtain an operating status time series correlation feature graph; Inputting the running state temporal correlation feature map into a feature attention selection and enhancement module based on a compression-inhibition structure to obtain an enhanced running state temporal correlation feature map; The global mean vector of the enhanced operating state time series correlation feature graph is calculated to obtain the enhanced operating state time series correlation feature vector.
3. The edge terminal security threat analysis and assessment method according to claim 2 is characterized in that: Performing full-time domain collaborative correlation analysis on the time series of the operating status data to obtain an operating status time series correlation feature graph includes: The time series of the operating status data is arranged into an operating status full-time domain collaborative correlation matrix according to the parameter sample dimension and the time dimension, and then the operating status time series pattern feature extractor based on the void convolutional neural network model is used to obtain the operating status time series correlation feature graph.
4. The edge terminal security threat analysis and assessment method according to claim 3 is characterized in that: Inputting the running state temporal correlation feature map into a feature attention selection and enhancement module based on a compression-suppression structure to obtain an enhanced running state temporal correlation feature map, including: Calculating the global mean of each feature matrix along the channel dimension of the running state time series associated feature graph to obtain a running state time series feature compression information representation vector; Performing multi-scale feature extraction on the running state time series feature compression information representation vector to obtain a multi-scale associated feature vector of the running state time series feature compression information; Using a Sigmoid function, a normalization operation is performed on the multi-scale associated feature vector of the running state time series feature compression information to obtain a feature weight vector representing the running state time series feature; Based on the running state time series feature representation feature weight vector, feature amplification and suppression operations are performed on the running state time series association feature graph to obtain the enhanced running state time series association feature graph.
5. The edge terminal security threat analysis and assessment method according to claim 4 is characterized in that: Performing multi-scale feature extraction on the running state time series feature compression information representation vector to obtain a multi-scale associated feature vector of the running state time series feature compression information, including: Performing one-dimensional convolution coding on the running state time series feature compression information representation vector to obtain a correlation representation feature vector between the running state time series feature compression information; Cascading the running state time series feature compression information representation vector and the running state time series feature compression information association representation feature vector to obtain the running state time series feature compression information multi-scale representation vector; The multi-scale representation vector of the compressed information of the operating status time series feature is input into a compressed information feature extraction module based on a multi-layer perceptron model to obtain the multi-scale associated feature vector of the compressed information of the operating status time series feature, wherein the multi-layer perceptron model includes two fully connected layers and a SiLU activation function.
6. The edge terminal security threat analysis and assessment method according to claim 5 is characterized in that: Based on the running state time series feature representation feature weight vector, the running state time series correlation feature graph is subjected to feature amplification and suppression operations to obtain the enhanced running state time series correlation feature graph, including: The enhanced running state time series association feature graph is obtained by calculating and multiplying the corresponding feature values of the feature weight vector representing the running state time series feature with the feature matrices along the channel dimension of the running state time series association feature graph.
7. The edge terminal security threat analysis and assessment method according to claim 6 is characterized in that: The time series of the network traffic value and the time series of the user behavior data are respectively subjected to time series association coding to obtain a network traffic local time series association implicit feature vector and a user behavior time series semantic coding feature vector, including: Inputting the time series of the network traffic value into a sequence encoder based on an RNN model to obtain a local temporal correlation implicit feature vector of the network traffic; The time series of the user behavior data is input into a temporal context encoder including a user behavior embedding encoder to obtain a temporal semantic encoding feature vector of the user behavior.
8. The edge terminal security threat analysis and assessment method according to claim 7 is characterized in that: Performing feature interaction correlation analysis on the enhanced operation state time series correlation feature vector, the network traffic local time series correlation implicit feature vector, and the user behavior time series semantic coding feature vector to obtain a security threat posterior time series representation vector, including: The enhanced operating state time series association feature vector, the network traffic local time series association implicit feature vector and the user behavior time series semantic coding feature vector are input into a security threat posterior analyzer based on a Bayesian probability network to obtain the security threat posterior time series representation vector.
9. The edge terminal security threat analysis and assessment method according to claim 8, characterized in that: Determining whether there is a security threat to the edge terminal based on the security threat posterior time series representation vector includes: The security threat posterior time series representation vector is input into a classifier-based threat assessment result generator to obtain an assessment result, and the assessment result is used to indicate whether there is a security threat.
Citation Information
Cited By
Network security multi-mode intelligent detection method and device, equipment and storage medium
CN120956524A
